Datasets:

Project-AgML
/

grape_variety_classification

dataset_info:
  features:
    - name: image
      dtype: image
    - name: label
      dtype:
        class_label:
          names:
            '0': Aswud Balad
            '1': Deas Al-Annz
            '2': Frinsi
            '3': Halawani
            '4': Kamali
            '5': Riasi
            '6': Shdah
            '7': Thompson Seedless
  splits:
    - name: train
      num_bytes: 2461447056
      num_examples: 8000
  download_size: 2094589734
  dataset_size: 2461447056
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-4.0
task_categories:
  - image-classification
size_categories:
  - 1K<n<10K

Grape Variety Classification

A dataset for variety classification of Grape leaves. The dataset contains 8,000 images across 8 classes:
Images per class:

Aswud Balad: 1,000
Deas Al-Annz: 1,000
Frinsi: 1,000
Halawani: 1,000
Kamali: 1,000
Riasi: 1,000
Shdah: 1,000
Thompson Seedless: 1,000

This dataset is indexed on https://project-agml.github.io/ as part of the AgML python library.

Citation

@article{al2023image,
  title={Image dataset of important grape varieties in the commercial and consumer market},
  author={Al-Khazraji, Lafta R and Mohammed, Mohammed Abdallazez and Abd, Dhafar Hamed and Khan, Wasiq and Khan, Bilal and Hussain, Abir Jaafar},
  journal={Data in Brief},
  volume={47},
  pages={108906},
  year={2023},
  publisher={Elsevier}
}

Mohammed Abdallazez, Mohammed; Al-Khazraji, Lafta Raheem Ali; hamed, dhafar; Hussain, Abir; Khan, Wasiq; Khan, Bilal (2022), “Grape Varieties Dataset”, Mendeley Data, V2, doi: 10.17632/7n3d6696hz.2

This dataset was reformatted from its original format to match HuggingFace standards.